Screening Of Biomarkers For Atrial Fibrillation Based On Metabolomics

Background:As one of the most common arrhythmias,atrial fibrillation(AF)often induces a variety of clinical complications.Atrial fibrillation can reduce heart function and,in severe cases,lead to thromboembolism or even death.Patients with atrial fibrillation has a case fatality rate twice that of the control group.The mechanism of occurrence and maintenance of atrial fibrillation is complex,which has not been solved in the field of arrhythmia.It has been reported that the recurrence and prognosis or onset of atrial fibrillation to be associated with inflammatory factors,blood clotting,etc.,but there is no specific examination could specifically diagnose or predict the prognosis of AF.Recent studies suggest that the main mechanism is atrial metabolic remodeling caused by atrial fibrillation,including disordered glucose metabolism,tricarboxylic acid circulation and enzyme dysfunction,which leads to the increase of oxidative stress products and the occurrence of atrial metabolic remodeling due to the uneven conduction of atrial electrical activity.Metabolomics aims to understand biological processes under normal or disease conditions by studying a complete set of small molecules in an organism.It can be used as a tool for the discovery of biomarkers and provide effective information for the mechanism and diagnosis of atrial fibrillation.Objectives:Based on liquid chromatography positive ion electrospray ionization tandem mass spectrometry,metabolomics analysis was used to explore new biomarkers for the diagnosis and prognosis of atrial fibrillation.Methods:1)Sample collection:plasma and atrial appendage samples were collected from patients with or without atrial fibrillation in Fuwai hospital of Chinese academy of medical sciences.Atrial appendage samples were collected during the operation following the surgical principle.The atrial appendage was cut into small pieces,transferred to a clean tube,and immediately stored at-80℃ until use.The blood samples were collected with an ethylenediaminetetraacetic acid(EDTA)anticoagulation tube the morning after the patients were hospitalized.2)Sample processing and metabolomics analysis:acetonitrile extraction method was used in plasma,2-step extraction method was used in atrial appendage samples,liquid chromatography positive ion electrospray ionization tandem mass spectrometry was used to detect the samples,and the original data files of metabolites in plasma and tissue samples were obtained.3)Data analysis:after data normalization,Simca-p 13.0 software was applied for multivariate statistical analysis(including principal component analysis,partial least squares analysis and orthogonal partial least squares analysis)to conduct principal component analysis for all samples.The VIP values of the first principal component of the orthogonal partial least square method and the student’s t-test(P<0.05)were combined to find the metabolites with differences.Then,the similarity between metabolites and samples is further studied through heat map analysis.The clustering software is R language.Correlation analysis was used to calculate the correlation coefficient between metabolites and metabolites.Pearson correlation coefficient was used as the method.Correlation matrix was calculated with cor()function of R language,and the correlation heat map of metabolites was drawn.The relationship between metabolites in the metabolic pathway construction was based on the KEGG metabolic pathway,and the enrichment analysis was carried out by metaboanalyst.The potential biomarkers were evaluated by ROC curve while differential metabolites were obtained.The analysis software was SPSS22.0.The sensitivity and specificity of biomarkers were evaluated by area under ROC curve(AUC).Results:A total of 165 patients were included in this study,including 49 patients in the atrial fibrillation group and 116 patients in the control group.A total of 60 cases of auricular tissue and plasma samples were collected from the two groups(atrial fibrillation group:30 cases;Control group:30 cases)and plasma sample:163 cases(atrial fibrillation group:48 cases;Control group:115 cases).262 and 361 metabolites for atrial appendage samples and plasma samples were acquired,respectively.Consequently,24 metabolites in atrial appendage samples and 24 metabolites in plasma samples could be mined for reflection of metabolic differences between AF and non-AF patients(VIP≥1,P≤0.05).Five identical metabolites including creatinine,D-glutamic acid,choline,hypoxanthine,and niacinamide(VIP≥1.5,P<0.01,FDR<0.05)in atrial appendage and plasma samples were considered prominent features of AF patients and D-glutamine and D-glutamate metabolism pathway was identified as the feature of AF patients.Finally,in plasma samples,the combinational markers comprised of D-glutamic acid,creatinine,and choline got a AUC value of 0.927(95%Cl:0.875-0.979,P<0.001),and displayed 90.5%sensitivity and 83,3%specificity were defined as the combinational biomarkers to recognize AF and non-AF.Conclusion:The results of this study suggest that metabolomics can be used as an effective tool to screen biomarkers.The combined biomarkers of D-glutamate,creatinine and choline can well predict the occurrence of atrial fibrillation.Changes in D-glutamine and D-glutamate metabolic pathways may be associated with underlying mechanisms of atrial fibrillation.